New biomarkers of drug-induced liver injury (DILI) are required in the clinic and in preclinical pharmaceutical evaluation. Liver-enriched microRNAs are promising serum biomarkers of acetaminophen-induced acute liver injury in mice. The utility of circulating microRNAs as biomarkers of human acute DILI is discussed in the context of correlation with existing biomarkers of liver injury and patient outcomes in acetaminophen toxicity, mechanisms of cellular microRNA release, and their potential advantages over current clinical biomarkers of DILI.

Toxicogenomics represents the merging of toxicology with technologies that have been developed, together with bioinformatics, to identify and quantify global gene expression changes. It represents a new paradigm in drug development and risk assessment, which promises to generate a wealth of information towards an increased understanding of the molecular mechanisms that lead to drug toxicity and efficacy, and of DNA polymorphisms responsible for individual susceptibility to toxicity. Gene expression profiling, through the use of DNA microarray and proteomic technologies will aid in establishing links between expression profiles, mode of action and traditional toxic endpoints. Such patterns of gene expression, or `molecular fingerprints' could be used as diagnostic or predictive markers of exposure, that is characteristic of a specific mechanism of induction of that toxic or efficacious effect. It is anticipated that toxicogenomics will be increasingly integrated into all phases of the drug development process particularly in mechanistic and predictive toxicology, and biomarker discovery. This review provides an overview of the expression profiling technologies applied in toxicogenomics, and discusses the promises as well as the future challenges of applying this discipline to the drug development process.

The goal of this study was to design a model system for the assessment of phototoxic potential using a human reconstructed epidermis (HRE, SkinEthic Laboratories, Nice, France), by testing some representative phototoxic (P) and non-phototoxic (NP) compounds and finished topical products. The tissue response to 24-h application of 5-5000 microg/mL of the test agents in the presence and absence of UVA light was analyzed in terms of viability (Lactate Dehydrogenase release), pro-inflammatory activity (IL-8 release and mRNA expression) and morphology (histopathology). 8-Methoxypsoralen (P) and promethazin (P), but not sodium lauryl sulfate (NP) produced cytotoxicity concentration-response curves significantly different between irradiated and nonirradiated tissues. Only irradiated tissues showed morphological damage. Application of tetracyclin (P) in the culture medium, but not topically, induced similar signs of phototoxicity. 6-Methylcoumarine (weak P) was not cytotoxic, yet it increased IL-8 release and mRNA expression only following UVA irradiation. PUVA therapy creams containing 1% 8-Methoxy-psoralen (P) or coal tar (P) decreased viability and induced histologic damage in UVA-exposed tissues. In conclusion, the phototoxic potential of the tested agents was correctly predicted by using a tiered strategy that involves determining cytotoxicity, production of IL-8, and morphological damage following exposure of the HRE to the compounds and UVA light.

Towards the end of their research and development efforts, pharmaceutical companies seek to deliver safe and efficient molecules. The availability of the human genome sequence shall contribute to reach this goal.

To the Editor Despite the loss of epidermal barrier function and subsequent exposure to environmental microorganisms, superficial epidermal wounds in vertebrates usually heal without major complications, suggesting the existence of a functional antimicrobial mechanism during wound healing. This general observation provides the basis for testing a hypothesis for the existence of a functional antimicrobial mechanism, which is active at cutaneous surfaces. By using standard immunohistochemical techniques, we have investigated expression of human P defensin-2 (hBD-2) in normal and lesional human skin. hBD-2 is a member of the defensin family of antimicrobial peptides, and has microbicidal activity to Gram negative bacteria and Candida albicans (Schroder and Harder, 1999). Moreover, hBD-2 has been shown to provide a link between innate and adaptive immunity (Yang et al, 1999). hBD-2 was first isolated from psoriatic scales (Harder et al, 1997), and is strongly expressed by psoriatic keratinocytes at sites of epidermal inflammation (Liu et al, 1998; Fig 1a). Although hBD-2 was not detected in the intact adult human epidermis (Fig 1b) or in the neonatal cornified foreskin epidermis (Fig 1c), it was also clearly detectable in the mucosal epithelium of the neonatal foreskin (Fig 1d), a finding that may well explain why mucosal surfaces rarely develop clinically relevant infections, despite the absence of a stratum corneum. Most notably, this study shows that hBD-2 is induced in the regenerating epidermis of acute surgical wounds (Fig 1e), suggesting that keratinocyte activation in response to loss of epidermal barrier function involves induction of an intrinsic antibiotic mechanism. Of special interest is the observation that hBD-2 is expressed in the epidermis of Apligraf (Fig 1f), a tissue- engineered human skin equivalent (HSE) produced from neonatal human foreskin. This product consisting of living fibroblasts anchored in a bovine type I collagen lattice and living keratinocytes forming a stratified epidermis has been developed for the treatment of wounds (Wilkins et al, 1994) and has beneficial effects in patients with chronic venous leg ulcers (Falanga et al, 1998). Unlike the other defensin genes, the hBD-2 gene contains several binding sites for the nuclear transcription factor NF-KB (Liu et al, 1998), which mediates responses to lipopolysaccharide (LPS) and proinflammatory cytokines. Therefore, we determined ifhBD- 2 expression is upregulated in the HSE following a topical bacterial challenge or exposure to LPS. Topical treatment with 0.5 p,g per ml (5 ng per cm2) and 1 p,g perml (10 ng per cm2) purified bacterial LPS (from E. coli serotype 0127:B8; Sigma-Aldrich, Bucks, Switzerland) for 24 h caused a strong, concentration-dependent, and statistically significant (p < 0.05) upregulation of hBD-2 expression (30.9 ± 14.6 and 59.1 ± 4.5; data represent fold change versus untreated HSE ± SEM). Recent studies have suggested that a close interaction between CD14 and TLR4 participates in LPS signaling, leading to nuclear translocation of NF-KB (Jiang et al 2000); however, to our knowledge, neither CD14 nor TLR-4 has been found to be expressed in keratinocytes. Therefore, a CD14/ TLR-4 independent LPS response mechanism may exist in keratinocytes. Alternatively, the possibility that a contaminant in the used LPS preparation might be responsible for the strong induction of hBD-2 in Apligraf cannot be excluded. Nevertheless, the hBD-2 induction in keratinocytes appears to be the result of pro-inflammatory processes, since other known chemical mediators of skin inflammation, e.g., phorbol 12-myristate 13-acetate and sodium lauryl sulfate, also enhanced hBD-2 mRNA expression in the HSE (data not shown). To investigate the effect of bacterial challenge on hBD-2 mRNA expression in Apliraf E. coli K12 bacteria grown over night in TY medium, pelleted, washed twice, and resuspended in DMEM/HAM-F12 (1:1) cell culture medium, were applied topically onto the epidermal surface of the HSE (approximately 106 bacteria per cm2). Bacterial challenge resulted in a weak but significant (p < 0.05) upregulation of hBD-2 mRNA expression (2.0 ± 0.4 after 24 h; 3.0 ± 0.7 after 48 h; data represent fold change versus untreated HSE ± SEM) as revealed by standard Real Time RT-PCR analysis. The observation that nonvirulent bacteria are much weaker inducers of hBD-2 than virulent bacteria, such as mucoid Pseudomonas aeruginosa (Harder et al, 2000), may provide a rationale for the fact that only a weak induction ofhBD-2 was achieved, although a rather high cell number of E. coli K12 bacteria was applied topically. To investigate if hBD-2 expression in the epidermis has functional relevance, we have soaked Apligraf with E. coli growth medium and subsequently exposed its surface to E. coli followed by a 24 h incubation at 37 °C. Bacterial colonies did not grow on the epidermal surface of the intact HSE, although colonies clearly formed on the dermal surface exposed to bacteria following mechanical separation of the epidermal layer. In conclusion, the findings of this study suggest that wound healing involves induction of an intrinsic antibiotic mechanism, which is also functional in epidermis of tissue-engineered skin. The authors thank Prof. Tomas Ganz, UCLA School of Medicine, Los Angeles, for kindly providing the anti-hBD2 antibody.

We have investigated the cytokine response pattern following sensitisation (induction) of BALB/c mice with different chemicals (dinitrochlorobenzene, dinitrofluorobenzene, oxazolone, glutaraldehyde, formaldehyde, trimellitic anhydride, croton oil) and elicitation (challenge) of contact allergy in sensitised animals. The results of our investigations showed that different chemicals induced both T helper (Th) 1 cytokines [interleukin (IL) 2, interferon β (IFNβ)] and Th2 cytokines (IL-4, IL-10) at different stages during murine contact allergy. We also confirmed our previous findings that IL-4 and IL-10 release were up-regulated during the challenge phase regardless the contact allergen used, whereas the release of IFNβ did not show a clear preference for being up- or down-regulated. In our hands, the increased expression of Th2 cytokines after challenge exposure to contact allergens appeared as a stable marker of secondary contact allergenic responses. Quantitative differences in the expression of IL-4 were observed between different contact allergens. The present results clearly indicate that skin sensitisers were able to elicit cytokine response patterns, which could not be related to a clear-cut Th1 or Th2 type of cytokine response. Furthermore, dermal application of contact allergens produced different kinetics of cytokine secretion upon induction and challenge. In our hands, the co-expression of Th1 and Th2 type cytokines appeared as a universal consequence of dermal application of contact allergens to responsive mice. Our results indicate that co-expression of Th1 and Th2 cytokines during contact allergy is an important feature of murine contact allergy in responsive mice and that chemicals differ in their potency to induce the expression of these cytokines. Furthermore, the results do not support the view that different chemicals induce Th1 or Th2 cytokines in a mutually exclusive manner depending on their preference for inducing either contact or respiratory allergy. The results are expected to renew the discussion about the usefulness of the Th1/Th2 paradigm in certain areas of immunotoxicology.

Cyclosporine A (CsA) is a potent immunosuppressant with the drawback of renal side effects. We reported that CsA markedly decreases calcium-binding protein calbindin-D28k mRNA levels in rat kidneys, and showed that this decrease is associated with its adverse renal effects. The transcription of the calbindin-D28k gene is activated via the vitamin D pathway. In this work, the potential CsA-mediated impairment of the vitamin D pathway was investigated. Wistar rats were treated for 12 days with 50 mg/kg/day CsA or for 20 days with 50 mg/kg/day of the non-immunosuppressant and non-nephrotoxic SDZ PSC 833, which had been previously shown not to affect calbindin-D28k mRNA levels. The expression of the three vitamin D-regulated genes calbindin-D28k, 1,25-dihydroxyvitamin D3-24-hydroxylase (24-OHase), and vitamin D receptor (VDR) were quantified in rat kidney homogenates by real-time reverse transcription-polymerase chain reaction. Plasma parathyroid hormone (PTH) as well as plasma and kidney 1,25 dihydroxyvitamin D3 (calcitriol) levels were monitored in all animals. CsA induced a 85% decrease in calbindin-D28k mRNA levels as well as a 40% and 69% decrease in VDR and 24-OHase mRNA levels, respectively. Plasma and kidney 1,25 dihydroxyvitamin D3 as well as plasma PTH levels were increased by CsA, but not by SDZ PSC 833. The treatment with SDZ PSC 833 did not affect calbindin-D28k or VDR expression, but did cause a 73% decrease in 24-OHase mRNA levels. Taken together, these results indicate an association between CsA-mediated down-regulation of rat renal calbindin-D28k mRNA and the decrease in other 1,25 dihydroxyvitamin D3-regulated genes, suggesting an impairment of the vitamin D pathway by CsA which may be related to its adverse renal side effects and its immunosuppressive activity.

Recently, we reported that in rat, cyclosporine A (CsA) markedly decreases the levels of calbindin-D (CABP-D) 28 kDa in kidney. CABP-D 28 kDa is a calcium-binding protein which is highly expressed in calcium-transporting tissues such as kidney or brain. In this study, we investigated whether, in addition to the kidney, CsA also has an effect on CABP-D 28 kDa in rat brain. Three groups of male Wistar rats received 15 mg/kg/day or 50 mg/kg/day of CsA orally for 12 days, whereas controls received vehicle solution for the same period. CABP-D 28-kDa protein and CsA were quantified in homogenates of kidney, cerebral cortex and cerebellum, and the localization of CABP-D 28 kDa was assessed in the different tissue sections by immunohistochemistry. In kidney, CABP-D 28 kDa was strongly and dose dependently decreased, and was located in tubular epithelial cells. In brain, CABP-D 28 kDa was not changed and was mainly located in pyramidal cells of the cortex and in cerebellum exclusively in Purkinje cells. High CsA concentrations were measured in kidney, more than 17-fold greater than those found in cortex. In cerebellum, CsA was below the limit of detection. These data suggest that at clinically relevant doses, CsA may not affect CABP-D 28-kDa levels in brain.

Using two-dimensional gel electrophoresis (2-DE), we recently discovered an association between decreased calcium-binding protein, calbindin-D 28 kDa, urinary calcium wasting and intratubular corticomedullary calcifications in rat kidney. This observation prompted us to investigate kidney tissues of other species, including man. In this paper we show that in dogs and monkeys, which are generally devoid of cyclosporine A (CsA)-mediated nephrotoxicity, renal calbindin levels were not affected by the CsA treatment whereas in CsA-treated human kidney-transplant recipients with renal vascular or tubular toxicity, a marked decrease in renal calbindin-D 28 kDa protein level was found in most of the kidney biopsy sections. The present results strongly suggest that calbindin is a marker for CsA-nephrotoxicity. The discovery of calbindin-D 28 kDa being involved in CsA toxicity has evolved from the application of 2-DE and has not been reported previously, proving that proteomics can provide essential information in mechanistic toxicology. Considering the current improvements in proteome methods it is expected that high throughput proteomics will become an indispensable tool in preclinical safety testing.

Cyclosporine A (CsA) is a potent immunosuppressant with the drawback of renal side-effects. We recently reported that relatively high doses of CsA markedly decreased the calcium-binding protein calbindin-D28kDa in kidneys of male Wistar rats, and showed that this decrease could be associated with some of the drug-induced adverse renal effects. To investigate the events leading to this decrease, the calbindin-D28kDa mRNA level in kidneys of rats treated with 15 or 50 mg/kg/day CsA for 12 days was analysed by reverse transcription followed by polymerase chain reaction. At both doses, a marked dose-dependent decrease in the calbindin-D28kDa mRNA level was found, one very similar to the decrease measured in the calbindin-D28kDa protein abundance. Thus, the CsA-mediated down-regulation of the renal calbindin-D28kDa protein is most likely the result of a decrease in the calbindin-D28kDa mRNA level.